逆向工程光栅投影点云模型配准技术的研究

Study on Point Cloud Registration of Raster Projection in Reverse Engineering

三维点云配准技术的核心是建立两个待配准点云之间的对应关系。根据协方差分析计算点云配准需要的旋转矩阵和平移向量,将目标点云转换到基准坐标系中实现点云配准。采用光栅投射式三维扫描仪获取某花瓶4组点云数据,利用提出的算法实现无序点云精确配准,并分析点云对之间的转换误差。实验结果表明:4片点云配准耗时约为6 s,配准的平均误差为0.027 mm,配准完成的点云轮廓完整平滑,达到反求曲面要求。该方法适合精度要求较高的自由曲面三维点云自动配准。

The key of 3D point cloud registration is to build corresponding relation between two pieces of point cloud.Rotating matrix and translation vector were obtained according to covariance analysis.Point cloud registration was implemented by transforming object point cloud to norm coordinates system.4 groups of vase point cloud were obtained by raster projection 3D scanning instrument,and unorganized point cloud registration based on proposed algorithm was realized.Transformation errors of point cloud couples were analyzed.The experimental results show that the registration time of 4 pieces of point cloud is about 6 s,and average error of registration is 0.027 mm.The point cloud profile after registration is complete and smooth,which can meet the need of surface reconstruction.The proposed method can be used in automatic registration of 3D point cloud of free surface.